This K02 Research Scientist Development Award (RSDA) application extends two methodologic approaches to be used in preventive trials in the mental health area. These approaches include the development of new designs as well as statistical models and data analytic procedures,. These new methods will be applied to both ongoing and new trials to prevent depression and conduct disorder. Work on this proposal will extend methods now being developed in an NIMH- funded R01 which involves for NIMH Prevention Centers (Arizona State, Johns Hopkins, Oregon Social Learning Center, and University of Michigan). These extensions will be based on new initiatives in prevention, as biologic, genetic, epidemiologic and developmental, etiologic, and treatment perspectives expand our opportunities for developing prevention trials. Linkages to key mental health researchers will be established to promote the principal investigator's collaboration in the next generation of biopsychosocial preventive trials. Three Internet-based methods will be used to provide efficient mechanisms for collaboration. Specifically, practical solutions for handling missing data, particularly attrition and selection bias, will be obtained and developed for prevention studies. A large class of non-linear and non-normal intent variable analyses will be developed using the Gibbs Sampler. New designs for preventive trials, especially those involving long-term follow-up will be developed and tested. These new methods will be applied to the analysis and design of trials examining the long-term impact on diagnoses of conduct disorder and depression. Other problems will address analytical issues arising from biogenetic studies an the use of biologic measures in prevention studies.